As comparative effectiveness research becomes a more prominent feature of clinical medicine, investigators and policy makers would do well to seek lessons from prior examples of this type of research. Our analysis of previous examples reveals lessons in three key areas: choice of comparison treatments, time frame of study, and widespread applicability of study results. Based on our observations, we offer suggestions for increasing the clinical applicability of comparative effectiveness research, such as employing surrogate endpoints that meet a specific threshold of validity. Future trials that address these areas of concern hold the greatest promise for improving patients' outcomes.